The development of hematologic malignancy in patients with hematopoietic failure is neither predictable nor easily treated. The hypothesis is that clonal cytogenetic abnormalities in the bone marrow are not predictive of evolution to leukemia, and that there may be other markers of myelodysplasia which do correlate with leukemia. This proposal is to determine the natural history, incidence, and evolution of clonal disease, myelodysplasia, and leukemia. The routine data which we plan to follow include clinical course, blood and bone marrow counts and morphologies. the investigative components include marrow chromosomes, in vitro hematopoiesis, cell cycle analyses, markers for aplastic anemia (activated T or NK cells), for myelodysplasia and leukemia, and for clonal disease (eg PNH), as well as genotype. We will study 25 to 50 patients with various bone marrow failure syndromes, including dyskeratosis congenita, Diamond-Blackfan anemia, Shwachman-Diamond syndrome, amegakaryocytic thrombocytopenia, and Kostmann's syndrome. We will particularly focus on Fanconi's anemia (FA), studying in particular 50 to 100 patients with FA who do not have leukemia at entry. This is the only condition where we anticipate the availability of a sufficiently large number of patients to perform statistical analyses. Patients with clonal marrow cytogenetic abnormalities will not be excluded. Treatment will not be regulated by this study, and patients will not be precluded from entering new treatment protocols. The power of this project is that all patients will be studied prospectively by the same laboratories. The significance is that we may identify early parameters which predict evolution to leukemia, and thus might impact therapeutic decisions.